542 
G. R. BISBY 
progressed, other considerations were forced upon the attention of 
the writer: as already noted, one has here to deal with closely related, 
variable hosts, and it is perhaps not strange that one should find the 
rusts also to be variable, and that the related American hosts should 
bear related rusts. In the course of the study here, between two and 
three hundred specimens of these autoecious rusts in the Arthur 
herbarium, assigned tentatively to different specific names, were 
examined. With the continuation of the work, it became more and 
more evident that the differences separating a few collections were 
neither of sufficient value nor constancy, as evidenced by an exami- 
nation of a large number of collections, to render possible a division of 
species upon a morphological basis. A few examples may aid in 
corroborating this view. 
The host Boisduvalia affords the most striking evidence of the 
variability of the rusts upon the Onagraceae. Holway, in his work 
upon the North American Uredineae, very logically described as a 
new species, Puccinia glabella, the rust occurring upon Boisduvalia 
glabella (Nutt.) Walp. This description was from a specimen dis- 
tributed by Griffiths, West American Fungi 385, as Puccinia Bois- 
duvaliae Peck. Holway points out that this rust upon Boisduvalia 
glabella is a very different thing from the rust occurring upon other 
species of Boisduvalia: Puccinia glabella has small teliospores (Holway 
gives 15-18 by 25-32 //), with the apex not at all or only slightly 
thickened; Puccinia Boisduvaliae possesses larger teliospores, usually 
much thickened at the apex. In a recent visit to the University of 
Wyoming, however. Dr. Arthur obtained further specimens of rusts 
upon Onagraceous hosts. Among these was a collection upon Bois- 
duvalia glabella showing somewhat larger teliospores, with thicker 
apices, than was shown by other specimens in the herbarium upon 
this host. The measurements of the teliospores of this specimen of 
Puccinia glabella were 15-19 by 26-35 M> the apex 4-7 ix. From the 
same herbarium a specimen was secured upon Boisduvalia densifiora 
(Lindl.) Wats., with spores but little larger, 16-23 by 32-39 ix, and 
the apex the same thickness, 4-7 11. Another collection in the her- 
barium upon Boisduvalia densifiora shows teliospores 18-26 by 32- 
53 II, with the apex 7-12 ju. That an error in identification of the 
host is not the explanation here, is indicated, not only by an examina- 
tion of the material, and a consideration of the carefulness of the col- 
lectors, but by the fact that other specimens of Boisduvalia rusts 
